French Researchers Demonstrate Superconducting Qubit With Zero Charge Drift for Three Months

Researchers at ENS Lyon and ENS Paris-Saclay have demonstrated a tantalum-based transmon qubit that maintains stable charge offset for nearly three months—a significant finding for a field plagued by unpredictable charge drift that disrupts qubit operation.

The team, led by Benjamin Huard (ENS Lyon Quantum Circuit Group) and Audrey Bienfait (ENS Paris-Saclay), reported in arXiv:2603.12367 that during measurements spanning approximately three months and two thermal cycles, the charge offset remained pinned at zero with no observable compromise to the qubit's coherence time (T1).

Charge offset drift has been a persistent practical challenge in superconducting qubits. Charges accumulating on the electrodes of a Josephson junction shift qubit energy levels unpredictably, limiting device performance and complicating scaling efforts. While transmons are designed to be less sensitive to charge noise than earlier qubit designs, they remain susceptible to slow drift in the charge offset parameter.

The stability the team observed was traced to an unexpected source: a thin superconducting layer that formed in parallel with the Josephson junction during fabrication, likely from incomplete wet-etching of tantalum on sapphire. This layer's inductance effectively pins the charge offset at zero.

However, the researchers caution that the mechanism is fragile—it disappeared in subsequent cooldowns, indicating the stability depends on specific, uncontrolled conditions during fabrication.

"Deliberately engineering such a layer offers a route to eliminating charge-offset drift in superconducting circuits more broadly," the authors note.

The work adds to growing evidence that tantalum-based qubits offer advantages in coherence. Earlier research demonstrated T1 times exceeding 0.3 milliseconds with tantalum, compared to 30-40 microseconds for conventional niobium-based transmons.

Assessment: This is a real experimental result with a clear practical implication. The three-month stability window is notable, though the fragility of the mechanism means reproducibility remains to be demonstrated. The proposed engineering solution—deliberately incorporating a superconducting shunt layer—could be testable by other groups. The result is not a demonstration of quantum advantage or error correction, but a meaningful improvement to qubit hardware that addresses a known failure mode.